Another Biffa TV22

There was a nice space behind the mains dropper to mount the transformer so I was able to keep the leads fairly short. I don't have good focus at present but the bandwidth looks pretty good. I remember Trevor mentioning about buffering. I wondered about putting some ferrites in the heater leads but I've got quite a bit to sort out before I get to that level.

Thanks,

Peter

Hi gents, it would be possible to wind a 1:1 transformer for 6.3 volts with quite low capacitance between the windings. This may actually work better than the normal mains powered version, it would also be smaller.

Ed

A bit like one of these http://www.radiomuseum.org/r/cbcelectro ... be_2f.html
Rob T

Thanks Ed, The one Alistair gave me has primary and secondary side by side. It looks promising.

Kind regards,

Peter

The transformer in question has a split bobbin so the capacitance should be reasonably low. It was made from a 20VA R.S. transformer kit that I wound my own secondary onto.

Having been rescued by Alistair's HKS transformer I've been trying to get past the stage I was at in the first photo in this thread but am now running out of ideas and looking for more expert advice. At the start of the thread I had a picture with low EHT (3.5kv) and poor line linearity and low width. Well, unfortunately I've not made much progress. My picture is as above with about 4.5kv EHT at min brightness, plenty of height but line linearity and width as shown.

I believe I've replaced all the leaky caps in the above circuit and I have checked the resistor values some of which were a little high.
I have good PL38 and PZ30. I tried replacing the Metrosil with a couple of diodes and a potential divider but this made zero change so I guess the Metrosil is OK.  I removed the LOPT and found that the primary (between pins 4 and 5) was low in inductance compared with G4JQT's measurements on various good TV22 LOPTs so Ed Dinning kindly rewound my primaries for me but on re-installing it I have slightly higher EHT but I still haven't found the fault.

I couldn't measure the overwind inductance but it runs totally cool. I then wondered if there might be a shorted turn in the scan coils but out of circuit the two halves show the same 7.3 mH inductance and 15 ohms resistance. I then started messing with the circuit values. If I reduce the value of C20 from its stated 470pF then I get slightly higher EHT but the linearity gets worse. The photo below is with 330pF. If I reduce R17 from 820k to 520k I get better linearity. If I reduce R21 from 470k to 330k I get better linearity. Anything more than minimum capacitance with TC1 reduces width and EHT. 

My HT is at 205 volts but my boost voltage is only 230volts. Nothing that I tried seems to give me sufficient EHT or boost voltage. The absolute highest EHT (at min brightness) I can get is 4.5kv. What do you think?

Peter

p.s. The frame distortion you see above is due to some 50 Hz ripple on the HT that I haven't bothered to sort.

I don't think the width coils are at fault as I kind of disconnected them. The linearity coils might be a problem. The service info does give resistances and mine measures ok but that probably doesn't indicate a shorted turn very well. What would be the best test for this?

Peter

I would have a go at getting rid of the mains ripple as it will also reduce the decoupling at the HT end of the LOPT and make diagnosing other faults more difficult.

The service info does give resistances and mine measures ok but that probably doesn't indicate a shorted turn very well. What would be the best test for this?

A measurement of Q?

Many years ago there was a project in Television to build a line transformer tester that worked on the principle that if you connected a defective LOPT with shorted turns to it, it damped the oscillator in the tester - thus stopping it - and gave an indication. Perhaps that's what we need here - the VRAT Patent Line Transformer Tester, one that doesn't need you to connect a LOPT to the calibrator output of a 'scope (something I've always been a little wary of, TBH) but gives a sure-fire indication of shorted turns.

I also have a Koenig diode-split LOPT tester that actually measures the output of said LOPT 

I would have a go at getting rid of the mains ripple as it will also reduce the decoupling at the HT end of the LOPT and make diagnosing other faults more difficult.

I take your point about poor decoupling but in fact I have new HT decouplers in place of the old leaky ones. I'm not sure why I have the ripple. There's no obvious overheating but perhaps something is drawing too much current somewhere but the Line Output has more than the rated HT voltage and should be OK for LF decoupling.

I disconnected the linearity coils and the C22 side measures 1.01 mH and 4.2 ohms and the other side 1.02 mH and 4.9 ohms. I also tried swapping the sides over and that results in an even more extreme linearity, totally compressed on the left.

LOPT Q measurement is probably the answer although the fault can only be in the EY51 heater winding or the EHT overwind. Perhaps Trevor's jelly pot route is the answer?

Peter

Remember that the TV22 was one of the first sets to employ a line timebase in which the line output valve was pulse driven rather than in earlier circuits where the line output valve was driven by a sawtooth waveform. In the TV22 the drive waveform turns the PL38 hard on and the current rises linearly in the LOPTx primary to produce the latter part of the scanning coils current. The efficiency diode PZ30 supplies the first 40% of the scanning waveform. It operates just like the line output stages of all modern TVs.
So if you consider the severe horizontal non-linearity of the raster, it is all too evident that the efficiency diode part of the line output stage is a fault. It certainly would be worthwhile checking the drive waveform at the control grid of the PL38.

Till Eulenspiegel.

Thanks Till,

Interestingly when I swapped the windings over on the linearity coils I got a totally cramped left side and even with me swapping them back it has remained totally cramped although strangely the winding resistances are little changed.

The PL38 grid looks like this at 20 volts/div and the flat region being 0volts

Peter

Hi Peter, there are some TV22 and TUG24 TVs in the shop and as the Biffa bin is full I can't dump them until Tuesday the day when they come to empty the bins here. When I return to the shop I'll compare the PL38 drive waveforms on one the TV22s. I'm sure it doesn't have to be a perfect sawtooth waveform.

Till Eulenspiegel.

Thanks Till,

I'm sure you've sent me on the right track directing me to the boost circuitry and the PL38 waveform makes more sense to me now given your description. I can't think what I've done to T5 to totally cramp the LHS but perhaps I should remove it and mess with its windings.

Peter

Hi Peter,

Have you checked or replaced the boost capacitor C21 (4mfd)? Also, it doesn't seem to matter too much about C22, if it is leaky or even short circuit the boost circuit will continue to work. The boost diode PZ30 and C21 are effectively shunted across the scan coils. I guess the operation of the boost circuit is as follows. The high energy flyback pulse present at the anode of the PZ30 charges C21 and during the interval when the PL38 is not conducting the charge in the capacitor flows into the scan coils thus creating the first part of the scanning stroke.

With regard to the PL38 control grid drive. The positive going tip of the waveform should be at near ground potential and the rest of the waveform is negative going. The grid1 and cathode of the valve works effectively as a clamp diode. The 0.01mfd coupling capacitor (C17) between the ECL80 oscillator and the grid of the PL38 should be checked.

Till Eulenspiegel.

Hi Till,

C21 is 2uF on my schematic but I didn't have a high voltage one of that value so I've got an 8uF in there but it doesn't appear too sensitive to value. I confess that I am still using the original C22. I tried series connecting a larger capacitor with it to eliminate any leakage problem but as you suggest C22 doesn't appear too critical.

My grid PL38 coupler C17 is new and correct value and the waveform is from 0 volts and negative going so I don't think that's the problem.

I'll be away from it for the rest of today but I'll have a hack around T5 again.

Thanks,

Peter

Hi Peter.
It looks to me like there is some damping after flyback or even during flyback. I know at least one person has had a shorted turn or turns in the linearity or width coil. Either could give you that effect.

Well I replaced the linearity coil with a couple of (uncoupled) 1mH inductors today and whilst the width was far too small the linearity looked quite good. So with this and that fact that messing with the linearity coil made it much worse I think Trevor's shorted turn suggestion probably hits the spot.

Looking at the coil there is a turn in the centre that looks suspiciously lacking in varnished insulation. Does anyone have a spare T5 they'd like to sell me?

Peter

That looks like a single layer coil, so shouldn't be too difficult to rewind.

I'd take an accurate measurement of the length of the winding, then remove the old one carefully, counting the turns as I go. Then it is an easy matter to divide the length by the number of turns to find the wire diameter.

You can then deduce the gauge from a suitable table but, unless you are lucky enough to find anyone with a stock of enamelled copper wire in SWG ranges, you'll probably have to opt for the nearest decimal size.

With care, you could probably replace the winding manually but, in the absence of a proper coil winder, an electric drill on a stand will do - it really needs to have a fairly slow speed setting, though. Rather than attempt to count the turns, just ensure that they go on close wound until you reach the correct winding length.

Thanks Terry,

My "linearity coil" description was perhaps a bit misleading. It's actually a more of a transformer, one to one with one winding on top of the other. But you are right. I think I have nothing to lose by attacking it although I might be mistaken in assuming I've spotted the short location. It could easily be in the bottom layer.

Peter